机器学习法分析两种透析质量指标持续达标评估体系对维持性血液透析患者预后的影响  

Machine learning algorithm of two continuous assessment methods of dialysis quality indicators based prediction scheme for assessing mortality risk in maintenance hemodialysis patients

作　　者：董建华 张明伟 杨茜 李川 吴边 孔凌 葛永纯 DONG Jianhua;ZHANG Mingwei;YANG Xi;LI Chuan;WU Bian;KONG Ling;GE Yongchun(National Clinical Research Center for Kidney Diseases,Jinling Hospital,Nanjing 210016,China;School of Health Sciences and Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]东部战区总医院国家肾脏疾病临床医学研究中心,南京210016 [2]上海理工大学健康科学与工程学院

出　　处：《肾脏病与透析肾移植杂志》2024年第4期301-307,共7页Chinese Journal of Nephrology,Dialysis & Transplantation

基　　金：江苏省自然科学基金面上项目(BK20201235)。

摘　　要：目的:利用机器学习法分析两种透析质量持续达标评估方法对维持性血液透析(HD)患者预后的影响。方法:筛选2016年1月在国家肾脏疾病临床医学研究中心接受HD治疗,且每年完成至少3次透析质量评估的患者。随访截止至2022年10月,终点事件为全因死亡。采用指标达标时长比和指标达标波动值作为9种透析质量指标(透析间期体重增长率、透析前收缩压、血红蛋白、血清白蛋白、血总二氧化碳、血钙、血磷、血全段甲状旁腺激素和单室尿素清除率)持续达标的评估方法,基于机器学习算法构建HD患者1年后存活或死亡的预测模型,并获得模型最佳概率阈值。结果:本队列研究共纳入240例HD患者,60例(25.0%)患者死亡。采用K-近邻算法(KNN)、随机森林(RandomForest)、极度随机树(ExtraTrees)、极限梯度提升树(XGBoost)、自适应增强(AdaBoost)和决策树(DecisionTree)六种机器学习法,分别构建基于透析质量指标达标时长比和指标达标波动值的预测模型。基于指标达标时长比的ExtraTrees模型具有最佳的预测效果,其准确率、精确率、召回率、F1分数和受试者工作曲线下面积分别达到0.92、0.86、0.96、0.91和0.93,同时证实0.65作为模型的最佳概率阈值。结论:基于透析质量指标达标时长比的机器学习模型对HD患者预后具有良好的预测效果。Objective:Use machine learning method to analyze the impact of two continuous assessment methods of dialysis quality indicators on the prognosis of maintenance hemodialysis(HD)patients.Methodology:A total of 240 patients who received HD treatment at the Eastern Theater Command General Hospital in January 2016 were screened,and dialysis quality was assessed more than three times a year.The follow-up period ends in October 2022,and the endpoint is death from all causes.The indicator time-to-standard ratio and indicator fluctuation value were used as the evaluation methods for the continuous achievement of nine dialysis quality indicators.Dialysis quality indicators include interdialytic weight gain、pre-dialysis systolic blood pressure、hemoglobin、albumin、total carbon dioxide、calcium、phosphorus、parathyroid hormone and spKt/V.A prediction model for survival or death of HD patients after 1 year was constructed based on a machine learning algorithm,and the optimal probability threshold of the model was obtained.Results:After 94 months of follow-up,60 patients(25.0%)died.Six machine learning methods,KNN,RandomForest,ExtraTrees,XGBoost,AdaBoost and DecisionTree,are used to build prediction models based on the indicator time-to-standard ratio and the indicator fluctuation value.The ExtraTrees model based on the indicator time-to-standard ratio has the best prediction effect,with its accuracy,precision,recall,F1 score and area under the receiver operating curve reaching 0.92,0.86,0.96,0.91 and 0.9 respectively,while confirming 0.65 as the optimal probability threshold for the model.Conclusion:The machine learning model based on the indicator time-to-standard ratio has a good prediction effect on the prognosis of HD patients.

关 键 词：血液透析 透析质量 机器学习 预后 

分 类 号：R692.5[医药卫生—泌尿科学]